We use utilities to backup databases. Setting up a regular database backup MS SQL Server Sql server backup write a script

View the base "Test _Recovery" at the camp " t4».

Let's start before updating the database from the new backup "Full2_Test_Recovery.bak" with SQL Server Management Studio ". Click with the right mouse button on the base " Test_Recovery ”, select “ Tasks", gave "Restore", then "Database".

At the vikni Restore Database" at the "Source" branch, select "Device". Dali « Add ”, write the way “\\ vniz-tst-bkp 01. test. local\Backup_SQL\Full 2_Test_Recovery. bak”, press “Ok”. At the "Destination" branch, select Database Test Recovery

Press "Ok"

The base is successfully renewed.

Let's take a look at the data base update, using Transact-SQL.

Right-click on the Test_Recovery database, select New Query in the menu:

At the vіkni we enter:

USE master

RESTORE DATABASE Test_Recovery

FROM DISK= "\\vniz-tst-bkp01.test.local\Backup_SQL\Full2_Test_Recovery.bak"

WITH REPLACE

The base is successfully renewed.

For which application, the REPLACE parameter was selected:

Revision does not allow the data base to be overwritten by another data base. If the database is already specified in the RESTORE instruction on the streaming server, and the GUID family identifier for the given database is revisited as the GUID family identifier for the database recorded in the backup dataset, then the renewal will not be discarded.

The REPLACE parameter is an example of important rewrites, as they sound like the renewal operation. There are such reverifications.

Rechecking for renewal over the main data base of the backup copy created for the other data base.If you change the REPLACE parameter for renewal, you can write data over the original data base independently, depending on how the data bases are located in the backup data set, and if the data is assigned to it, it will be considered as recorded in the backup set. It is possible to bring to vipadic overwriting over the data base of another data base.
Revalidation of data bases, such as the replacement model of the overhaul or the model of the restoration with incorrect protocols, for which the backup copy of the final fragment of the log was not created, and the STOPAT parameter was not stopped.If the REPLACE parameter is not correct, it is possible to overwrite the data that was recorded, the remaining pieces of data written to the log have not yet been copied to the backup copy.
Overwriting current files.

The great functionality of Bacula Enterprise Edition, apart from that, allows you to quickly and easily create backups of the database. For example, go about the tool, for the help of which you can create a backup copy of MS SQL Server. It is possible to make backups of MS SQL backups by creating backup copies of specific MS SQL databases of the great ones that are broken by the Windows platform, with smaller amounts of data on third-party compilers, with the possibility of updating data up to the first time (PITR-local backup) on a disk drive.

The Bacula Systems script for creating backups of MS SQL Server is characterized by superior efficiency, which reaches for the implementation of a modern, high-performance architecture. Moreover, the software allows you to create a backup of MS SQL Server, to achieve the most powerful ability to create backup copies of MS SQL.

Backup script for MS SQL Bacula Systems works independently from VSS. This means that MS SQL Backup Tool does not use VSS snapshots to create backups. To that end, you can set the value “Enable VSS = no” to the Bacula FileSet. Efficiently creating backups of MS SQL Server and their support for additional solutions are available for the Microsoft API for SQL Server. In order to do this, Bacula Systems can support the robot's security mechanisms and all types of authenticity implemented in Microsoft SQL Server.

Backing up the MS SQL transaction log and updating MS SQL at the time of the hour: The Bacula Enterprise Edition software allows you to restore blocks of MS SQL data, or specifically update to the next time. For implementations of models of complete renewal and incomprehensible protocols, you can implement MS SQL, PITR-renewal or LSN to update the system to a specific location. You can restore the MS SQL data base to any moment of the hour to the exact second. At different backups of the MS SQL transaction log, when the database is updated, it will be updated from different selected backups.

Brief overview of functionsautomatic backup and update of MS SQL from Bacula Enterprise

Bacula Systems has created a backup plugin for MS SQL Server for the Bacula Enterprise Edition. Backup MS SQL Server from Bacula can have the following functions:

Support for a new and differential backup of MS SQL
MS SQL Incremental Backup Guide
Backup MS SQL on a partition and a local disk
Backup copy of MS SQL for distribution
Creation of backups on equal databases of MS SQL Server
Ability to enable/disable the database from the procedure for creating backups
Support for creating backups of the database "only for reading"
Update MS SQL backups to disk
Enforcing the backup stream without intermediary in the Storage Daemon
Update MS SQL at the moment

Review and setup of MS SQL backup 2008, 2008 R2, 2012 and 2014

This document has solutions for Bacula Enterprise Edition 8.4 and later versions that are not supported by earlier versions of the software. MS SQL database backup was tested and supported by MS SQL 2003 R2, MS SQL 2008 R2, MS SQL 2012, MS SQL 2005, MS SQL 2008, MS SQL 2014. MS SQL backup robot from Bacula from SQL Express is possible.

MS SQL Backup Glossary 2008, 2008 R2, 2012 and 2014

MS SQL stands for Microsoft SQL Server.
Transaction log. Whether the MS SQL Server database has a transaction log, before which all transactions and modifications of the database are recorded, which lasts for an hour of such transactions. The transaction log is an important element of the database. In different systems, the transaction log may be needed to upgrade the database to a working state. You can find detailed information at https://msdn.microsoft.com/en-us/library/ms190925.aspx.
Differential backup of MS SQL Server databases. Differential backup of foundations for the rest of the day. At the end of the differential backup, these data are stored, as if they were changed from the moment the remaining full backup was created. You can find detailed information at https://msdn.microsoft.com/en-us/library/ms175526.aspx.
External backup of MS SQL Server databases. As a result of a new database backup, a backup copy of all data bases is created. Backup includes part of the transaction log with a method of updating the database again from a backup copy. Renew DB backup Restore the DB at the time of completion of the backup copy. You can find detailed information at https://msdn.microsoft.com/en-us/library/ms186289.aspx.
Backup "only for copying" (CopyOnly). Backups “only copies” are MS SQL backups, independent of the obvious sequence of creating traditional SQL Server backups. Sometimes you can create backups for special needs, without interfering with the process of backing up and updating the database. You can find detailed information at https://msdn.microsoft.com/en-us/library/ms191495.aspx.
VDI(Virtual Appliance Interface) - all Microsoft technology that allows you to create naming channel between programs.
standard masks are set and rows of substitution characters. For example, the standard mask production* includes rows production1 and production2.
row
whole number.
LSN A skin entry in the MS SQL Server transaction log is assigned in addition to a unique transaction registration number (LSN). More detailed information can be found at https://technet.microsoft.com/en-us/library/ms190411%28v=sql.105%29.aspx.

Backup MS SQL Server 2008, 2008 R2, 2012 and 2014

External backup of MS SQL Server 2008, 2008 R2, 2012 and 2014 databases

As a result of a new backup of the MS SQL database, the database files and the transaction log are saved, which allows you to restore the MS SQL database from time to time. At the time of one or more files of the MS SQL database database backup, allow to restore all transactions. Also, you will be notified of the successful transactions that were changed in the process of winning. In this mode, the backups of the master and mbdb databases are created.

Differential backup of MS SQL Server 2008, 2008 R2, 2012 and 2014 databases

Differential backup of MS SQL Server databases on the remaining full backup of MS SQL databases. During the creation of a differential backup of MS SQL, those data are stored, as if they were changed from the moment of creation of the remaining full backup of MS SQL. For the function of differential backup of MS SQL, the sequence of backups is also important. For some reason, a new backup, which MS SQL uses, is not available, differential backups of the MS SQL Server database cannot be beaten. Backup copying of MS SQL in Bacula vicorist method and solution of the problem. Therefore, in times of vindication, the status of a differential database backup can be automatically promoted to a full backup.

Backup copy of transaction log MS SQL 2008, 2008 R2, 2012 and 2014

Setting up MS SQL backup and database configuration

Update MS SQL database from backup

You can use all standard methods to launch the procedure for updating the MS SQL database from a backup. However, you are to blame for the change in the fact that at the time of updating the differential data, the latest forward backup of the MS SQL database will also be updated. At this time, the update will automatically run, so you run yoga in the console bconsole for additional update options 5 or 12. In order to generate file structures, it is necessary to specify the update of the latest databases or database instances.

Options for restoring MS SQL databases from backup

The Bacula Enterprise Edition software allows hackers to set up impersonal options for updating MS SQL and to set up different ways to “check in” the database. Renovation options, which are most often vicorated, are described below:

Where parameter: For the Bacula Enterprise Edition dropdown, this parameter allows the administrator to update the database in a specific location.
Replace parameter: Changed to determine how Bacula can behave from a threaded database when it is backed up. MS SQL backup like Bacula also allows you to add more options when upgrading, for example:
Instance: MS SQL oscillators victorious copy of instances, MS SQL database backup like Bacula allows you to choose which instance should be restored. This parameter is optional, i, even though there are no tasks, when updating the selection, the value specified when creating the backup is taken. For the promotion, the instance named “MSSQLSERVER” is awarded.
database. This option specifies the name of the database to be updated and the value set at the time of the creation of the database. This parameter is optional. When locking backups of SQL Server databases, select the Where parameter to specify the name of the new database. If the Where and Database parameters are set to a valid DB name, the Database parameter will be invalidated.
user. I'm a coristuvacha, like a vicorist for connecting to an instance of MS SQL data base. This parameter is optional, i, even though there are no tasks, when updating the selection, the value specified when creating the backup is taken.
password. The password to be used to connect to the MS SQL database instance. This parameter is optional, i, even though there are no tasks, when updating the selection, the value specified when creating the backup is taken.
domain. Domain, which is selected for connection to MS SQL data base instance. This parameter is optional, i, even though there are no tasks, when updating the selection, the value specified when creating the backup is taken.
recovery. The parameter allows you to set whether the database will be brought to the front when it is updated. For zamovchuvannyam when updating the database, I’ll make a cut to the front.
Stop_before_mark Umov WITH STOPBEFOREMARK = Vikoristovuetsya in order to show that the record in the journal of transactions, which is known directly before the ensign, is the point of inspiration. The point of origin can be date and hour, LSN or mark_name.
Stop_at_mark. Umov WITH STOPATMARK = Ticked in order to show that the transaction is designated as the point of origin. STOPATMARK moves forward to the flag, which includes retrying the marked transaction. The point of origin can be date and hour, LSN or mark_name.
Stop_at= . Umova WITH STOPAT = vykoristovuetsya in order to indicate that the point of origin is the date/hour.
Restrict_user. Umov's WITH RESTRICT_USER wins for sharing access to the updated data base. For promotions, the value of no.

Updates to MS SQL at this time can be downloaded directly from the MS SQL backup plugin. You can also check files locally and check operations from the Microsoft SQL Server Mangement Console, so that you can change more than you can.

LSN

The LSN number of the entry in the journal, at the time of the particular winkle, the specific time the backup was created, can be looked over in one of the following ways:

When I see a description of the task of creating a backup for additional software Bacula
Name the log file
Table msdb.backupset
Table msdb.backupfile

When the task is not created, the backup of the MS SQL database, when the description of the task is displayed, will show up information about the LSN numbers:

Number First LSN send the remaining LSN number of the remaining backup to the transaction log. This backup can be either the first new backup or the remaining backup (incremental).

Number Last LSN confirm the rest of the transaction registered in the journal.

In the case of a backup to the transaction log (incremental), naming the file associated with the DB chain, in the task of creating an incremental backup, it will look like an offensive rank:

The number in the name, in our mind 42000162001, Reply to the remaining LSN number of the previous backup (when creating a new or incremental backup).

Figure 2: First LSN, Remaining LSN and Filename LSN

As shown in the application for little 2, as the administrator needs to update the MS SQL database in the country, which confirms LSN number 14, you can use the following:

At the database update menu, select option 5
Select the rest of the backup file “data.bak” (LSN: 10)
Select incremental backup “log-10.trn”

Otherwise, if the last backup of MS SQL Server is not available, but the front backup is available, then:

Victory update option 3, wrap the default values of the jobids
Select database directory “/@mssql/db29187”
Choose a new backup file “data.bak” (LSN: 2)
Select incremental backups "log-2.trn", "log-3.trn", "log-10.trn"
Set the stop_at_mark parameter equal to “lsn:14”
Start backup task

MS SQL upgrade scenarios

Description	Where	Database	Butt
Upload files to disk	way		where=c:/tmp
Check out the database			where=/
Innovate with new names	Im'ya		where=newdb
Innovate with new names		Im'ya	database=newdb
Update with new names and move files		Im'ya

Table 1: MS SQL Upgrade Scenarios

2.3.1 Updating MS SQL databases from different names

To restore the database with the removed names, parameter Where not responsible for assignments (empty value), otherwise you can set the value “/”, and the parameter Replace may be important Always, otherwise it is necessary to remove the database from the beginning.

Renewal of MS SQL backup with new names

In order to restore a backup of the MS SQL data base with new names, it may be necessary to move the database files to disk. Keep everything in view of the fact that the DB is still out.

If the database is not available, then the parameter where, or the “Plugin Options” field can be named the new data base. Backup MS SQL in Bacula automatically create a database from new names.

If the database is still needed, the where parameter will be changed to move the files to disk, and you will need to specify the name of the new database behind the “Plugin Options” menu. You need to select the file layout.dat for the tree of innovation.

Vicorist My Catalog

Run MS SQL update task:

After selecting the My Catalogue, run the MS SQL database update task:

Update MS SQL to a local drive

How to say where=c:/path/, the files will be updated to the local disk, and the MS SQL database administrator can change the TSQL extension procedure for the Microsoft SQL Server Management Console to update the database. SQL commands, database updates required, recaptured in the description job output as shown below.

2. Revision of information about backup and renewal for the database

In order to recognize, if the creation of backup copies of a specific data base was created, as well as the restoration of the data base from a backup copy, you can speed it up with a standard sound " » (Backup and Restore Events). To form this sound, in the Server Objects browser, right-click on the input data base, select " Call» (Reports) - « Standard sound» (Standard Reports) - « Submit backup and update» (Backup and Restore Events).

Formation of the following data:

Average Time Taken For Backup Operations
Successful Backup Operations
Backup Operation Errors
Successful Restore Operations

To review the information, it is necessary to expand the grouping of the stars.

Chi helped you chia article?

Learn how to copy tables from MS SQL Server databases. Proponu dekіlka options for creating copies of the table. Which one of them to choose - to lie in the structure of the tables, to appear in the new indexes, to trigger too thinly, and also to work with your hands.

1. Manual method of copying the table structure

In Microsoft SQL Management Studio, select the database, select the table, right click on the mouse and select "Script Table as" -> "CREATE TO" -> "New Query Editor Window". At the request, the code of the table creation is displayed. It is necessary for the new one to specify the name of the base, it is necessary to make a copy of the table, and the new name, so the base is not changed. How to create the code for creating the table structure is shown in the little box below.

In addition to this method, an index of tables will be created, but triggers will not be accumulated. It is necessary to copy them in a similar way.

To copy the data from the already created table, it is necessary to extract the following SQL query:

INSERT into ..tmp_tbl_Deps SELECT * FROM ..tbl_Deps

2. Copying SQL tables in one row

Make a copy of the structure of the tables and data in the middle of one database:

SELECT * into tmp_tbl_Dep FROM tbl_Deps

Copy the structure of the table and її data from one base to another:

SELECT * into ..tmp_tbl_Deps FROM ..tbl_Deps

The downside of this solution is that indexes are not copied.

Irrespective of those that in the previous materials, we already discussed the backup copying of Microsoft SQL Server databases, a reader's opinion showing the need to create a comprehensive material with more deep-seated theoretical part. True, vikonans with an emphasis on practical statements of the article allow you to quickly improve the backup copy, but do not explain the reasons for choosing quieter ones. We will try to fix this clearing.

Models of renovation

The first step is to set up a backup copy, then select the renewal model. For an optimal choice, the next step is to evaluate the extent to which the criticality of the data wasted, making them superimposed on the implementation of this other model.

As you can see, the MS SQL database is composed of two parts: the data base itself and the transaction log before it. The database to retrieve the data and service data at the current moment of the hour, the transaction log includes the history of all changes in the data base for the song period, with the transaction log we can change the base to be a certain moment of the hour.

For sampling, two models of innovation are introduced in the sorrel media: it's simple. Use also model inaccurate protocols, but it is recommended only as an addition to the new model for the period of large-scale mass operations, if it is not necessary to update the base at the right moment of the hour.

simple model transferring a backup copy of only the database data, we can only update the database at the time of the backup copy creation, all changes in the interim between the creation of the remaining backup copy and the failure will be lost. At the same time, the scheme is simple with small invoices: you need to save only a copy of the data base, the transaction log is automatically truncated and does not grow at the expansions. Also, the renewal process is the simplest and does not take a lot of time.

Povna model allows you to update the database at any given time, but also, save backup copies of the database, save a copy of the transaction log for the entire period, for which you may need a backup. With active work with the database, the expansion of the transaction log, also, and the expansion of archives, can reach greater expansions. The renewal process is also richly folded and truncated by the hour.

When choosing a renewal model, follow the credit for renewal with the credit for taking backup copies, and also take into account the qualifications and qualifications of the staff, which will lead to renewal. In addition, with a new model, the knowledge will be sufficient for the staff of the first qualification, even if the scheme is simple, it will be enough to follow the instructions.

For databases with a small amount of data to be added, it may be more convenient to win a simple model with a high frequency of copies, so that you can quickly update and continue the work by entering the entered data manually. It is the fault of the model in the first line that it is guilty of victoriousness there, the waste of data is unacceptable, and it is possible to reinforce it with significant vitras.

See the backups

New copy of database- as a clear name, in the data base, that part of the active transaction log for the hour that the backup copy was formed (to keep track of all ongoing transactions). Allows you to restore the database at the time of the backup.

A copy of the data base- Even a copy may be one stottny nedolіk, won't avenge all the information of the data base. As backup copies need to be done frequently, due to the power of the non-economous disk space, so much of the collection will be taken up by the same data. For usunennya tsgogo nedolіku can vikoristovuvat retail copies of the data base, yakі stavіt іnformatsiyu, scho chіnіlasa hіѕ hіѕ hіѕ hаvе аnd thе аnd copying.

Zvertaemo respect, retail copy - tse danі vіd the moment of the rest again copy, tobto. dermal advancing retail copy to avenge the data in front (but if it smells, it can be changed) and the size of the copy will be constantly growing. For renewal, one copy and one retail copy is enough, call the rest. The number of retail copies should be chosen according to the growth of their expansion, as only a few retail copies are compared with the expansion of half of the total, it is possible to create a new copy.

Backup copy of the transaction log- zastosovuetsya only for the new model of renewal and to replace a copy of the transaction log starting from the moment the previous copy was created.

It is important to remember the coming moment - copies of the transaction log are not related to copies of the data base and do not erase the information of previous copies, so for updating the base you need to have an uninterrupted copy of copies of that period, so you want to be able to be basic. At any moment of the remaining successful copying may be in the middle of the period.

Marveling at the little one, as if the first copy of the log file will be used, then you can only restore the base station at the time of the new copy, which will be similar to the simple model of renewal, restore the base station at any moment of the hour, you can only a little after the offensive repetitive) copying, For the mind, that the lancets of copies of journals starting from the front copying of the base and far away will be uninterrupted (for a little one - from the third and far).

Transaction log

To understand the processes of renewal and recognition of different types of backup copies, see in more detail the attachments to the transaction log. Transaction - the least possible logical operation, as it may be sensible and may be more victorious. Such a pіdkhіd zabezpechuє tsіlіsnіst і nesuperechnіst danih for any situations, oskolki promіzhny stan operation is unacceptable. To control whether there are any changes in the destination database, the transaction log.

In case of any operation, a record about the transaction start is added to the transaction log, a unique number (LSN) is assigned to the skin record from an inconsequential sequence, in case of a change in these data, a log entry is made, and if the log is recognized, the completion of the operation (fixation) transactions.

When triggered on the skin, the system analyzes the transaction log and reads all uncommitted transactions, changes roll forward at once, if they are recorded in the log, or not written to disk. Tse gives the opportunity to win back the cache and deposit records, not being afraid for the integrity of data to wind for the availability of backup live systems.

That part of the journal, as a way to avenge active transactions and win for the renewal of data, is called the active part of the journal. It is based on a number, which is called the minimum renewal number (MinLSN).

In the simplest form, MinLSN is the record number of the first incomplete transaction. If you look at the little ones more, then when you see the blue transaction, we take MinLSN equal to 321, after fixing in record 324, the MinLSN number changes to 323, which is consistent with the green transaction number that has not yet been fixed.

In practice, it's a little more complicated, for example, these closed blue transactions can still not be thrown to disk and the MinLSN relocation to 323 can be restored to the invalid operation. In order to avoid such situations, the concept of a checkpoint was introduced. The control point is created automatically for the advancing minds:

At the time of the explicit CHECKPOINT instructions. Control point spratsovuє at the streaming data base of the day.
In case of a larger data base operation with minimal registration, for example, in case of a larger data base copying operation, the renewal model with inaccurate protocols expands.
An hour to add or delete files in databases, using the ALTER DATABASE instruction.
When running an instance of SQL Server for the additional SHUTDOWN instruction, or when running a SQL Server (MSSQLSERVER) service. In that case, and in another way, a skin database checkpoint will be created in the instance of SQL Server.
This is how the SQL Server instance periodically creates automatic checkpoints in the skin database for a quick update of the database.
Under the hour of creating a backup copy of the data base.
Under the hour of vikonannya diї, which means the inclusion of the data base. Stocks can be set to the AUTO_CLOSE parameter ON and close the rest of the session with the database or change the database parameter, which will restart the database.

Regardless of how the date was earlier, MinLSN will assign a checkpoint record number value, or the oldest pending transaction.

Transaction log enhancement

A transaction journal, like a journal, requires periodic cleaning of old records, otherwise the loan will grow all the time. Vrahovyuchi, scho with active work with the database of transaction logs can significantly change the size of the database, the network is currently rich in administrators.

Physically, the transaction log file is a container for virtual logs, which are sequentially filled in from the log growth. The logical log, which should replace the MinLSN record, is the top of the active log, the front logical logs are inactive and are not needed for automatic database update.

If a simple renewal model is selected, then when the logical logs reach the equal 70% of the physical file, the inactive part of the log, sov. reduction. However, do not bring the log before changing the physical file, the logs are truncated only logically, as after the operation they can be beaten again.

If the number of transactions is large and at the time of reaching 70% of the physical file's size, there will be no inactive logical logs, then the size of the physical file will be increased.

In this way, the transaction log file with a simple update model will expand to the activity of the work with the doty base, the docks will not be able to accommodate the entire active part of the log. Pіslya what yogo growth to stick to.

With the new model, the inactive part of the journal cannot be used up, the docks cannot be used up until the backup copy. Update the log to be carried out for the sake of knowing that a backup copy of the transaction log was made, after which the checkpoint was created.

Incorrect configuration of the transaction log backup under the new model can lead to uncontrolled log file growth, which is often a problem for unsupervised administrators. It is also often used to manually amplify the transaction log. With a new model of renewal, this work is not categorical, so that you destroy the integrity of the copies of the magazine and you can restore the database only at the time of the creation of the copies, so you can confirm a simple model.

It’s time for you to guess for yourself, about what we were talking about on the cob of the article, so that you can turn the model over to the other side to reinvent it, to see the difference between the simple model.

A simple innovation model

Now, after acquiring the necessary minimum knowledge, we can move on to a detailed review of inspiration models. Let's start from idle time. Suppose, at the time of the battle, we had one line and two retail copies:

The backup copy was made once for production and the remaining copy was created at night from the 21st to the 22nd. Zbіy v_dbuvaєtsya in the evening of the 22nd before the creation of the cherg's copy. At the same time, we will need to renew the rest of the retail copies, with which we will spend the rest of the working day. For whatever reason, the copy on the 21st will also appear late, then we can renew the front copy, having spent a day of work, however, if the copy for the 20th is not successful, it will not be possible to successfully renew the date on the evening of the 21st, if it is obvious copies.

Renewal model

Let's take a look at a similar situation, but from the standpoint of a new model of innovation. Backup copies are also sparse with us, according to the principle of cost + retail, as well as a copy of the transaction log for extraction.

The renewal process will be foldable at times. We need to manually create a backup copy of the final fragment of the journal (shown in chervonim), tobto. part of the journal from the moment the last copy was made before the accident.

If you don't create anything, then you can restore the database until I'm at the time of the creation of the remaining copy of the transaction log.

With every possible file of a copy of the journal for the previous day, do not start restoring the actual base camp, but surround us with the moment of creating the rest of the copy, tobto. streaming doby.

Then we will subsequently update the same retail copy and lancet copies of the magazine, created after the rest of the backup copy, the remaining copy of the final fragment of the magazine, which will give us the opportunity to restore the database right at the time of the accident, or enough to blow it.

If the rest of the retail copy will be saved, then for a simple model you will produce up to one working day, the same model allows you to redo the copy, after which you will need to redeem the entire copy of the transaction log until the moment you redo the copy. Glybina vіdnovlennya lie down only in vіd gіbіn uninterrupted lancet logіv.

On the other hand, if one of the copies of the transaction log is squandered, let's say, redistributed, then it is possible to restore the data only at the time of the remaining backup copy + period for the non-shackled lancet copies of the logs. For example, if magazines were scrambled in 12, 14 and 16 years and later a magazine, creations in 14 years, then it is possible to make a copy of the database until the end of the uninterrupted lance, tobto. Up to 12 years old.